This project is designed to examine the role of serotonin uptake inhibition as a pharmacologic treatment for the prevention of relapse to alcoholism. On the basis of prior animal and human studies, fluoxetine, a specific serotonin uptake inhibitor, appears to be a potentially useful therapeutic agent, in the maintenance of abstinence in alcohol dependent persons. The successful treatment of alcoholism is clearly of enormous health and social benefit both to the individual, and to society as a whole. The hypothesis is that serotonin uptake inhibition decreases the desire to consume alcohol. This study will specifically examine whether fluoxetine is a useful short term pharmacologic adjunct in the treatment of alcoholism. A randomized placebo controlled clinical trial will be conducted in 200 recently abstinent alcohol dependent persons, who will receive 60 mg of fluoxetine or placebo for a duration of eight weeks. The majority of alcohol dependent persons who relapse to drinking do so during the first few month following detoxification. Therefore pharmacologic (as well as other) treatment is likely to have greatest impact during this time period. The study population will include significant numbers of women and minorities (African-Americans). All subjects will concurrently be enrolled in intensive outpatient treatment which includes three visits per week with breathalyzer monitoring, counseling and group therapy, for a duration of six months. Subjects will complete daily diaries on alcohol consumption which will be collected at research visits weekly for the first two months, and then at two and a half, three and six months. Other compliance measures for abstinence will be determined at research visits, and will include urine alcohol and drug screens, laboratory tests (AST, ALT, GGT, MCV), and corrobration of self report from a collateral. Compliance with the taking of medication will be confirmed with pill counts and serum fluoxetine and norfluoxetine levels. Primary outcome endpoints will be established a priori and will include the proportion of abstinent subjects by group at one, and two months. Other analyses will include number of days of abstinence, days to relapse, and surrogate measures of drinking behavior such as laboratory tests and social stability. An intention to treat analysis will be conducted, and interim analyses will be allowed provided that the group are not identified. This study will provide important information on the utility of fluoxetine as a term pharmacologic treatment for relapse prevention.